← All companies
Mistral AI
European AI lab (Paris). Maker of the Mistral models and Le Chat Pro. The reference European foundation-model player; benefits from EU sovereign-AI demand and government partnerships. Employees mostly hold French BSPCEs.
Reviewed May 10, 2026 (6d ago)
Current estimate
4 methods · May 10, 2026Medium confidence (40%)
$14.0B
range $10.5B – $17.5B⚠ methods disagree (5.7× spread)
Implied per-share: $70.0 · vs last primary +17%
Last primary round
$12B
Series C · 11mo ago
Secondary signal
$13.0B
Hiive · 2026-03+8%
Headcount
500
employees · LinkedIn+79%
Revenue (est.)
$150M
annual run rate
Complete company data
Every input we use, where we got it from, and how much we trust it.
| Field | Value | Confidence | |
|---|---|---|---|
| Company fundamentals | |||
| Sector | AI | high | |
| Headquarters | 🇫🇷 France | high | |
| Founded | 2023 | high | |
| Website | mistral.ai | high | |
| Latest primary round | |||
| Post-money valuation | $12B | high | |
| Round name | Series C | high | |
| Date | 2025-06-04 (11 months ago) | high | |
| Lead investor | General Catalyst | high | |
| Amount raised | $900M | medium | |
| Capitalization | |||
| Total fully-diluted shares | 200M | low | |
| Implied share price (latest primary) | $70 | medium | |
| Primary equity type granted | ISO/NSO | high | |
| Financials | |||
| Annual revenue (est.) | $150M | medium | |
| EV / Revenue (implied) | 80.0x | medium | |
| Headcount | |||
| Employees (now) | 500 | medium | |
| Employees (12 months ago) | 280 | medium | |
| 12-month headcount growth | +79% | medium | |
| Methodology inputs | |||
| Sector ETF proxy | BOTZ | medium | |
| ETF annual return (trailing 3-yr) | 32% | medium | |
Funding history
Primary rounds reported in Bloomberg, WSJ, TechCrunch, Reuters, or SEC Form D filings (efts.sec.gov). Linked per round below.
Series C2025-06-04
$12.0BLed by General Catalyst · raised $900M
Source: Les Echos Jun 2025
Series B2024-06-11
$6.2BLed by General Catalyst · raised $640M
Source: TechCrunch
Series A2023-12-11
$2.0BLed by Andreessen · raised $415M
Source: Bloomberg
Secondary market signals
Secondary market signals: bid/ask/trade indications from Hiive (hiive.com), Forge Global, Caplight Index, and tender prices reported in the press.
| Date | Type | Source | Price / share |
|---|---|---|---|
| 2026-03-15 | trade | Hiive | $65 |
| 2025-12-02 | trade | Forge | $58 |
| 2025-09-05 | bid | Hiive | $48 |
View live indications: Hiive ↗ · SEC Form D filings ↗
Public peer comparables
Public peer comparables: editorial selection of public companies with the closest business model. EV/Revenue and growth from public 10-Q filings and Bloomberg, refreshed quarterly.
| Ticker | Company | EV / Revenue |
|---|---|---|
| NVDA | Nvidia | 22x |
| PLTR | Palantir | 55x |
| CRWD | CrowdStrike | 18x |
| MDB | MongoDB | 11x |
| NET | Cloudflare | 14x |
Typical strike prices by cohort
Strike ranges by cohort: aggregated from Glassdoor, Teamblind threads, Carta's published industry data and reports from employees. Indicative only — your actual 409A at grant time is the only authoritative number.
2023 cohort
$1–$3
per share
2024 cohort
$10–$20
per share
2025 cohort
$35–$50
per share
Confidence breakdown
How much we trust each of the four valuation methods for this specific company, and why.
Method A — Peer-multiple
weight 35%→ $3.0B23% confidence
↳ Effectively excluded from the blend — this method is structurally unreliable for this company.
Growth-weighted peer EV/Revenue: 25.5x → 22% private-company discount → 19.9x on $150M revenue. Output implies less than half the recent market anchor ($13B) — this company is likely valued on optionality / non-revenue factors and Method A is structurally less reliable here.
Method B — Secondary-implied
weight 25%→ $12.2B55% confidence
Recency-weighted average of 3 secondary indications (180-day decay)
Method C — Primary time-decay
weight 30%→ $16.9B66% confidence
$12B round (2025-06-04) × 2.84× growth × 0.95× multiple decay over 0.9 years — capped at 1.3× recent secondary ($13.0B → $16.9B)
Method D — Sector momentum
weight 10%→ $15.6B46% confidence
$12B × (1 + 32% p.j.)^0.9yr = $15.6B via BOTZ ETF-proxy
Overall confidence
40%
Weighted average of the four methods. Above 65% is high, 40–65% medium, below 40% low.
Worked example for an employee
A concrete walk-through using a hypothetical employee profile. Same engine the calculator uses — try it with your real numbers.
Profile: a hypothetical employee who joined Mistral AI 24 months ago (2024-05-16) with 2,000 options at a strike price of $15/share, on the standard 4-year vesting schedule with a 1-year cliff.
Vesting
1,000 of 2,000
50% vested at month 24
Implied share price (mid)
$70
weighted average of 4 methods
Gross vested value
$70,039
vested shares × implied price
Exercise cost
$15,000
1,000 × $15 strike
Net of exercise
$55,039
gross − exercise cost
Tax (rough)
$7,162
US federal only · AMT risk: low
What if the next round is…
Up round +30%
$64,175
New round at +30% valuation, 15% new shares issued
Down round −30%
$25,856
New round at −30% valuation, 20% new shares (higher dilution in distress)
Secondary today
$44,533
Sale on the secondary market at the typical 15% discount to the mid estimate
These numbers come from the same engine as the live calculator. Walk through with your own grant: use your real numbers →
Transparent assumptions
What we assumed where exact data isn't available. If any of these don't match your situation, the numbers above will be off.
Fully-diluted share count
200M shares
We don't have authoritative cap-table access for private companies. The number above is reconstructed from SEC Form D filings, secondary-market price-vs-valuation math, and leaked share-class summaries. Actual fully-diluted count (including unvested option pool and SAFEs) may differ by 5–15%.
Common-stock waterfall
1.0× to common
We assume a clean common-stock outcome — i.e. preferred-share liquidation preferences and ratchets do not extract value before common. In a down-exit, this assumption can be very wrong: preferred stack can take 30–70% of proceeds.
Private-company illiquidity discount
22% off public peers
Applied to Method A (peer-multiple). Per-company because category leaders with deep secondary markets and a clear IPO path warrant a smaller discount than early-stage names. The literature range is 10–35%; for Mistral AI we use 22% based on secondary-market depth, valuation history, and proximity to liquidity.
Revenue growth assumption (Method C)
200% YoY
Compounded forward from the last primary date to today to estimate the company's value drift. Set per-company based on disclosed or reported ARR trajectory. Method C is further capped at 1.3× the most recent secondary trade (if any) to stop runaway extrapolation when revenue growth outpaces what the market is actually paying.
Sector ETF momentum (Method D)
BOTZ @ 32%/year
Method D applies the trailing return of BOTZ as a market-proxy adjustment to the primary valuation. Crude — meant as a tiebreaker when peer and secondary data conflict, not as a primary signal.
Outlier auto-downweight
Methods >50% from median get ½ weight
Robust-statistics safety net: any method whose output strays more than 50% from the median of the other three methods has its confidence weight halved when computing the blended mid-point. Prevents a single bad extrapolation from dragging the headline.
Recency decay for secondary trades
180-day half-life (e⁻ᵈ/¹⁸⁰)
A trade from 30 days ago counts ~5× more than one from 180 days ago. Tunable; some platforms use shorter half-lives.
Tax estimate
US federal only, single filer
Uses 2024 US federal brackets, ignoring state tax, NIIT, AMT credit recovery, and any other deductions. Treat as a directional sanity check, not a tax bill.
All data from public sources. Not investment, tax, or legal advice. Not affiliated with Mistral AI. Last refreshed manually — see methodology for our update cadence.